Your autonomous engineering fleet, orchestrated.
Lumpy spins up, supervises, and self-heals fleets of Claude Code agents across your projects and servers — with an always-on Conductor keeping everything healthy around the clock.
Teams building on Lumpy
One platform to run your whole AI engineering operation
Lumpy turns a pile of one-off AI coding sessions into a coordinated, supervised, self-healing fleet.
Orchestrate Claude Code agents
Spin up autonomous coding sessions per project, give each one a task, and let them run. Start, stop, resume, and steer every agent from one place.
Fleet & uptime monitoring
Track every server, machine, and remote with live CPU, memory, disk and load — plus hosted-service uptime checks and alerts when something goes sideways.
Self-healing deploys
Ship changes with build, type-check and health verification — and automatic rollback if anything fails. Agents can safely improve the platform itself without taking it down.
The Conductor
A top-level orchestrator watches platform health and the whole fleet 24/7 — triaging problems, keeping sessions productive, and logging the state of everything.
Micro-services & Librarians
Deploy reusable specialist agents — like per-project Librarians that build living knowledge bases, or a Platform Health Auditor that keeps watch.
Drive it from anywhere
Approve a prompt, answer a question, or kick off work straight from your phone. Your fleet keeps moving whether you're at your desk or not.
From repos to a running fleet in minutes
Point Lumpy at your work, define what each agent should do, and let the Conductor keep it all healthy.
Connect
Link your projects, repositories, and servers. Lumpy imports them and maps your fleet automatically.
Spin up agents
Create autonomous Claude Code sessions per project with a task and the access they need.
Supervise
The Conductor monitors health, surfaces alerts, and keeps every session doing useful work.
Ship & self-heal
Changes deploy with verification and auto-rollback, so your platform stays up no matter what.
Teams that run more software than people can babysit
Whether you operate one product or a portfolio, Lumpy gives you a coordinated AI engineering layer that never sleeps.
Portfolio & agency teams
Run a dedicated agent fleet per client project, each with its own knowledge base and guardrails.
Platform & infra teams
Continuous monitoring, triage, and safe self-healing deploys across your whole server fleet.
Solo founders & small teams
Punch above your headcount — keep maintenance, tests, and upkeep moving without a full ops team.
Loved by the teams it runs for
Placeholder testimonials — swap for real quotes & attribution
“[ Customer quote about how Lumpy replaced a pile of manual babysitting with a fleet that just runs. ]”
“[ Customer quote about uptime / self-healing saving a weekend. ]”
“[ Customer quote about shipping more with a small team. ]”
Start small, scale your fleet
Placeholder tiers & prices — confirm packaging before launch
Solo
- 1 project, up to 2 agents
- Fleet monitoring & alerts
- Community support
Team
- Unlimited projects & agents
- The Conductor + self-healing deploys
- Librarians & micro-services
- Priority support
Enterprise
- SSO, roles & audit logs
- Self-hosted / VPC deploy
- Dedicated Nublear support
Put your engineering on autopilot
Join the waitlist for early access to Lumpy — your autonomous engineering fleet, orchestrated.
Questions, answered
What exactly is Lumpy? +
Lumpy is an orchestration platform for AI engineering. It runs and supervises fleets of Claude Code agents across your projects and servers, monitors the health of everything, and ships changes with self-healing, auto-rollback deploys — all coordinated by an always-on Conductor.
Do I need to babysit the agents? +
No. That's the point. Agents run autonomously on the tasks you give them, and the Conductor keeps the fleet healthy and productive. You step in only when you want to — and you can do that from your phone.
How do self-healing deploys work? +
Changes go through build, type-check, and health verification before they go live. If anything fails or the platform doesn't come back healthy, Lumpy automatically rolls back to the last known-good version — so a bad change never takes you down.
What can it monitor? +
Servers, machines, and remotes (CPU, memory, disk, load), plus hosted-service uptime checks. When something crosses a threshold or goes offline, Lumpy raises an alert and the Conductor triages it.
Is my code and infrastructure secure? +
Lumpy is built by Nublear with least-privilege access and scoped tokens, and Enterprise plans support self-hosted / VPC deployment so your code never leaves your environment. [Placeholder — confirm final security/compliance details before launch.]
When can I use it? +
Lumpy is rolling out in batches. Join the waitlist and we'll send an early-access invite.